JP2897082B2 - Motion compensation variable length coding - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a high-efficiency coding system for moving pictures, and more particularly to a motion-compensating variable-length coding system using a plurality of motion-compensating variable-length coding tables.

In a high-efficiency video coding method, information obtained by coding a difference between an original picture and a predicted value and a motion vector obtained by performing motion detection are each subjected to variable-length coding.
Generate the required variable length encoded output.

[0003] Such a motion vector encoding method is as follows.
It is desired that more efficient encoding can be performed by selecting an optimal table from a plurality of motion compensation encoding tables.

[0004]

2. Description of the Related Art FIG. 13 shows a conventional encoding circuit, in which reference numeral 101 denotes an encoder for encoding a difference between an original picture and a predicted value, and reference numeral 102 denotes variable-length encoding for an encoded signal. VLC unit 103, a variable length coding (VLC) unit 103 for a motion vector, 104, a frame memory for storing an image of one frame, 105, a variable delay unit, 106, a motion detection unit for performing motion detection, and 107, a subtraction unit. , 108 are adders. The variable delay unit 105 and the motion detection unit 106 form a motion compensation unit 109, and the VLC unit 1 for the encoded signal
02, the VLC unit 103 for the motion vector forms a variable length encoder 110.

[0005] The input signal is input in blocks of a plurality of pixels by block scanning of the original image signal, and is applied to a subtraction unit 107 and a motion compensation unit 109. In subtraction section 107, the difference (prediction error) between the signal of the original image and the signal of the predicted value from variable delay device 105.
Ask for. The encoder 101 encodes the difference signal and inputs the difference signal to the variable-length encoder 110 and the adder 108. The encoding is performed by, for example, vector quantization or discrete cosine transform (DCT). The adder 108 decodes the encoded signal and adds the signal of the predicted value to generate a reproduced signal. This signal is stored in the frame memory 104,
A signal of a reproduced image one coded frame before is generated. On the other hand, in the motion compensating unit 109, the motion detecting unit 106 detects a motion vector by matching the reproduced image of the previous encoded frame from the frame memory 104 with the original image. The detected motion vector is stored in the variable-length encoder 1
10 and a variable delay 105. In the variable delay device 105, the frame memory 10
By delaying the image signal of the pre-encoded frame from step 4, the signal of the aforementioned predicted value is generated.

In a variable-length encoder 110, a VLC unit 102 for a coded signal performs variable-length coding on an encoded difference signal, and a VLC unit 103 for a motion vector.
Then, a variable-length encoded output is generated by performing variable-length encoding on the detected motion vector. The VLC unit 103 for the motion vector has a motion compensation coding table, and performs variable length coding on the motion vector by outputting a codeword corresponding to the input motion detection result with reference to the table.

FIG. 14 is a diagram for explaining the matching of a motion vector. Reference numeral 11 denotes an original image, and reference numeral 12 denotes a reproduced image (previous image) one coded frame before in a frame memory. As a general method of matching for motion compensation, when the size of the search range for motion compensation is, for example, (+ x to -x, + y to -y), the original image block (a, b) 13 For the previous image 1
2, the difference between the original image block 13 and the image block in the previous image 12 is calculated within the search range (a + x to ax, b + y to by) indicated by 14 determined by the motion compensation table, and the accumulated difference is calculated. Assuming that the block with the smallest error is 15, its position (a + k, b + m)
Is determined as the motion vector 16. The difference in this case is often calculated by the absolute value error, and sometimes by the square error.

[0008]

In the conventional encoding circuit shown in FIG. 13, the characteristics of an input image differ from those of an image such as an image with a small motion or an image with a large motion. Even in such a case, only one type of variable length coding table for motion vectors is used. Therefore, for example, in the case of an image having an extremely large motion, there is a problem in that the motion compensation range of the encoding table is small for the motion, so that mismatching is likely to occur and image quality is deteriorated.

For an image having a small motion, the coding range of the motion vector in the coding table is unnecessarily wide. In a coding table having a wide motion vector coding range, a motion vector corresponding to a large motion is stored. Since a code word is included, a longer code word is allocated than necessary, which causes a problem that coding efficiency is reduced and improvement in image quality is prevented.

An object of the present invention is to solve such a problem of the prior art. In a high-efficiency moving picture coding method, a plurality of motion compensation variable length coding tables are provided. By using the information on the motion vector before the frame to select the optimal coding table according to the magnitude and nature of the motion and coding the motion vector, vector mismatching can be reduced and efficient motion can be achieved. It is an object of the present invention to provide a motion-compensated variable-length coding system capable of coding a vector.

[0011]

SUMMARY OF THE INVENTION According to the present invention, a coded signal obtained by coding a difference between a reproduced image one coded frame before and a predicted value is variable-length coded, and a motion detection unit performs one coded frame coding. In a high-efficiency moving picture coding method in which a motion vector detected by performing matching between a previous reproduced picture and an original picture is subjected to variable-length coding using a motion compensation coding table, a plurality of motion compensation coding tables and A memory for holding the motion vector of the previous one encoded frame, a calculation / judgment unit for judging the nature of the frequency of occurrence of the motion vector from the held motion vector of the previous encoded frame, And a switching unit for selecting the motion compensation coding table.

[0012]

In the high-efficiency coding method for a moving image, a coded signal obtained by coding a difference between a reproduced image one coded frame before and a prediction value is variable-length coded, and
By performing matching between the reproduced image one encoded frame before and the original image to detect a motion vector, and performing variable length encoding on the motion vector using the motion compensation encoding table,
Encode a moving image. At this time, the high-efficiency video coding scheme, a plurality of motion compensation coding table 2 ₁ -
2 _n is provided, the motion vector of the previous one encoded frame is stored in the memory 3, and the calculation / determination unit 4 is provided, and the occurrence frequency of the motion vector is calculated from the motion vector of the previous one encoded frame stored in the memory 3. The information is generated by judging the nature of the information. And according to this information in the switching unit 5 _1, 5 _2, performs switching to select a plurality of motion compensation coding table 2 ₁ to 2 _n.

In this case, the calculation / determination unit 4 determines whether or not the number of motion vectors that become 0 in one frame exceeds a certain threshold value, and the maximum and minimum values of the motion vectors in one screen. value depending on the whether the judgment result exceeds the respective threshold, it generates information for selecting a plurality of motion compensation coding table 2 ₁ to 2 _n.

In this case, the calculation / determination unit 4
The determination result as to whether the number of motion vectors having the maximum value or the minimum value in one frame has exceeded each threshold value, and whether the number of motion vectors having zero in one screen has exceeded a certain threshold value depending on the whether the judgment result, generates information for selecting a plurality of motion compensation coding table 2 ₁ to 2 _n.

Further, in this case, the calculation / judgment unit 4 obtains an average value of the motion vectors in one frame, and outputs information for selecting a plurality of motion compensation coding tables 2 ₁ to 2 _n according to the average value. Occur.

Therefore, according to the present invention, an optimum coding table can be selected from a plurality of motion compensation coding tables by using information of a motion vector one coded frame before. Less
Efficient encoding can be performed.

[0017]

FIG. 2 shows the configuration of an embodiment of the present invention, in which the same components as those in FIG. 13 are denoted by the same reference numerals, and 201 _{1 to} 201 _n are motion compensation coding tables 1 to 1, respectively. n. 202, 203 are switches, 2
04 is a memory for storing one frame of motion vectors, 20
Reference numeral 5 denotes a calculation / judgment unit for calculating and judging the magnitude, nature, etc. of the motion of the motion vector.

The input signal is input in units of blocks by block-scanning the original image signal.
07 and the motion compensation unit 109. Subtraction unit 107
, A difference (prediction error) between the signal of the original image and the signal of the prediction value from the variable delay unit 105 is obtained, the encoder 101 encodes the difference signal, and the variable length encoder 110 and the adder 108 Add to The adder 108 decodes the encoded signal, adds the decoded signal to the signal of the predicted value, creates a reproduced signal, and stores the reproduced signal in the frame memory 104 to generate a reproduced image signal. In the motion compensating unit 109, the motion detecting unit 106 matches the reproduced image of the previous encoded frame from the frame memory 104 with the original image, detects a motion vector, and outputs a variable length encoder 110 and a variable delay To the vessel 105. The variable delay unit 105 delays the image signal of the previous frame from the frame memory 104 in accordance with the motion vector, and generates a signal of a predicted value. The above operation is the same as that of the conventional encoding circuit shown in FIG.

In the variable-length encoder 110, the VLC unit 102 for the encoded signal performs variable-length encoding on the encoded difference signal and outputs the result. The result of the motion vector detection by the motion detection unit 106 is
Frames are stored. The calculation / determination unit 205 determines the magnitude, nature, and the like of the motion from the motion vector information of the pre-encoded frame stored in the memory 204, and gives the result to the switches 202 and 203.
Switching is performed so as to select any _one of the motion compensation coding tables 201 _{1 to} 201 _n , and the information of the determination result is fed back to the motion detection unit 106,
The search range at the time of motion detection is changed according to the calculation / judgment result so that the motion compensation coding table selection matches the operation range at the time of motion detection operation. Motion detector 1
The motion vector for the original image detected at 06 is subjected to variable-length coding using the selected motion compensation coding table.

As described above, in the motion-compensated variable-length coding system of the present invention, the optimum motion-compensated coding table can be selected according to the motion in the original image, so that the mismatch at the time of motion detection is reduced and the image quality is improved. And efficient coding can be performed. The receiving side uses the motion compensation coding table when decoding the transmitted variable length code, but the receiving side also uses the information of the motion vector coded one coded frame before like the transmitting side. Therefore, if the motion compensation coding table is selected and decoding is performed, the information of the motion vector one coded frame before has already been sent to the receiving side, so that no new information is sent.
Since the decoding can be performed by selecting the motion compensation coding table, the method of the present invention can be applied only by adding a few circuits to the conventional method. Hereinafter, a simple code word will be described as an example for easy understanding. in this case,
The range of the motion vector is ± 4-7.

3 to 6 show examples (a) to (d) of the motion compensation coding table according to the present invention, respectively. FIG. 3 shows that the motion is large but concentrated around 0. FIG. 4 shows the case where the movement is small but not concentrated very much near 0, FIG. 5 shows the case where the movement is large but not very concentrated around 0, and FIG. This shows a case where the user is concentrated in the vicinity.

FIGS. 7 and 8 are flow charts (1) and (2) showing an operation example in the embodiment shown in FIG. In particular, an example of the determination operation is shown.

First, in step 301, motion vector information VX (u, v) and VY (u, v) one frame before the encoded frame.
And threshold values TH1, TH2, and TH3 for determination. Here, u and v are the number of blocks in one frame. In step 302, an initial value is set. In the drawing, "/" represents X or Y. MINSUM · indicates the minimum value of the vector in the X or Y direction in one frame, and is set to any value larger than the range of motion vectors (search range) as an initial value. MAXSUM · indicates the maximum value of the vector in the X or Y direction in one frame, and is set to an arbitrary value smaller than the motion vector range (search range) as an initial value. COUNT · indicates the number of vectors that take a value of 0 for X or Y in one frame, and the initial value is set to 0. COUNTS indicates the number of motion compensations (the number of motion vectors), and the initial value is set to 0.

Next, with respect to the vector for one screen, the number of motion vectors (COUNTS) is counted for each of the vector VX in the X direction and the vector VY in the Y direction, and the minimum value of the vector (MINSUM ·) is obtained. The maximum value (MAXSUM ·) is obtained, and the number of vector 0 (CO
UNT) is counted (step 303). In the following steps 304 to 306, the approximate tendency of the motion vector occurrence frequency in the vectors VX and VY is estimated from the obtained maximum value and minimum value of the vector and the number of the vectors 0. If there is no motion vector, as in the case where the intra-frame encoding is forcibly performed, the calculation and counting of the maximum value and the minimum value are not performed. As a threshold value for determining the number of vectors 0 (COUNT ·) in step 304, a predetermined threshold value TH1 · is a ratio of the number of motion vector occurrences (COUNTS) to the total number of blocks (u × v). Is used.

In steps 304 to 306, the type of occurrence frequency is estimated using the threshold value. For example, if in step 304 the number of vectors 0 exceeds a certain threshold, and in step 306 the minimum value is less than the threshold value (TH2 •) and the maximum value exceeds the threshold value (TH3 •),
This is a case where a large number of vectors are concentrated around 0 and some large vectors exist. In this case, if the motion compensation coding table (a) shown in FIG. 3 is selected, the coding efficiency is good, and the process proceeds to step 307.

If the number of vectors 0 exceeds a certain threshold value in step 304 and the maximum value and the minimum value do not exceed the threshold value in step 306, the concentration is close to 0 and no large vector exists. It is. In this case, FIG.
When the motion compensation coding table (d) shown in (1) is selected, the coding efficiency is good, and the process proceeds to step 309.

If the number of vectors 0 is equal to or less than a certain threshold value in step 304, and if the minimum value and the maximum value exceed the threshold values in step 305, there is a large vector that is not concentrated near 0. This is the case. In this case, if the motion compensation coding table (c) shown in FIG.

If the number of vectors 0 is equal to or less than a certain threshold value in step 304, and if the minimum value and the maximum value do not exceed the threshold values in step 305, a large vector is not concentrated near 0 and exists. If not. In this case, if the motion compensation coding table (b) shown in FIG. 4 is selected, the coding efficiency is good.

In the operation example shown in the flow charts of FIGS. 7 and 8, the information input to the memory 204 includes "the range of the motion vector corresponds to the motion compensation coding tables 201 ₁ to 201 2.
01 _n must be possible up to the maximum possible value ". That is, by selecting the motion compensation coding table detected one coded frame before, ± 7
However, if the value can only be taken up to 4, it is restricted even if a vector with a larger value actually occurs, and a table with a larger vector should be selected after the next frame. Can not be done.

FIGS. 9 and 10 are flowcharts showing another operation example of the embodiment shown in FIG. 2, which is larger than the vector range in the motion compensation encoding table detected one encoding frame before. An example in which a motion compensation coding table having a larger vector can be selected after the next encoded frame when the vector of Table (b) shown in FIG. 4 or FIG.
Or (d).

Steps 401 to 403 are the same as steps 301 to 303 in the examples of FIGS. However, in step 402, for the X direction and the Y direction,
The number of vectors having the maximum value within the range of the motion vector (C
MAX.) And the number of vectors taking the minimum value (CMIN)
・) The initial value of is set to 0.

In step 404, the number of vectors having the maximum value within the range of the motion vector (CMAX ·),
Similarly, the number of vectors having the minimum value (CMIN ·) is checked. In step 405, the number of vectors having the maximum value (CMAX ·) checked in step 404 exceeds the threshold value (TH2), or the number of vectors having the minimum value (CMIN ·) exceeds the threshold value (TH3). Check if it is less than. Further, in steps 406 and 409, it is checked whether or not the number of the vector 0 (COUNT.multidot.) Exceeds a certain threshold value. The threshold for determining the number of vectors 0 in this case is the same as that shown in FIG.

If the number of vectors having the maximum value or the minimum value within the range of the motion vector exceeds the threshold value in step 405, and the vector values are concentrated near 0 in step 406, If the motion compensation coding table (a) is selected, the coding efficiency is high.
Go to 07.

If the number of vectors having the maximum value or the minimum value within the range of the motion vector exceeds the threshold value in step 405 and the vector values are not concentrated near 0 in step 406, If the motion compensation coding table (c) is selected, the coding efficiency is high.
Proceed to 08.

If the number of vectors having the maximum value or the minimum value within the range of the motion vector is less than or equal to the threshold value in step 405, and the vector values are concentrated around 0 in step 409, FIG. When the motion compensation coding table (d) is selected, the coding efficiency is high.
Proceed to 10.

If the number of vectors having the maximum value or the minimum value within the range of the motion vector is equal to or smaller than the threshold value in step 405 and the vector values are not concentrated near 0 in step 409, FIG. If the motion compensation coding table (b) is selected, the coding efficiency is high.
Go to 11.

In the operation examples shown in the flowcharts of FIGS. 7 and 8 and FIGS. 9 and 10, the information held in the memory 204 includes "0 and 0 Around ”is required. That is, the estimated average value of the motion vector is regarded as 0, the occurrence frequency of 0 is counted, and the highest frequency is determined. In particular, the motion compensation coding tables shown in FIGS. 3 and 6 are determined as having the highest occurrence of 0. However, if panning is taken into account,
The frequency of occurrence in the panning direction is the highest.

FIG. 11 is a flow chart showing still another example of the operation in the embodiment shown in FIG. 2, in which the smallest code word is assigned to the vector having the highest frequency of occurrence. In this example, it is considered that there is no improvement effect when there is not much concentration in the vicinity where the occurrence frequency is high, and the case where the concentration is high in the vicinity where the occurrence frequency is high is taken up.

FIG. 12 shows an example (e) of the motion compensation coding table according to the present invention, which is an example of the motion compensation coding table when the smallest codeword is assigned to the vector having the highest frequency of occurrence. Is shown.

In FIG. 11, step 501 is the same as in the operation example of FIG. 3 or FIGS. However, in this case, no threshold is required. In step 502, the average value AVR of the vector is obtained.
Are selected from the plurality of motion compensation coding tables 201 _{1 to} 201 _n so that α becomes the calculated average value AVR in the motion compensation coding table shown in FIG.

[0041]

As described above, according to the present invention, an optimum coding table can be selected from a plurality of motion compensation coding tables by using information of a motion vector one coded frame before. , And efficient encoding can be performed with less mismatch of vectors. Further, since the motion compensation coding table is selected using the information of the motion vector one coded frame before as the existing information,
On the receiving side, the motion compensation coding table can be selected and decoded in the same manner, and there is no need to newly transmit side information for table selection. The method of the present invention can be applied only by performing the method.

[Brief description of the drawings]

FIG. 1 is a diagram showing a basic configuration of the present invention.

FIG. 2 is a diagram showing one embodiment of the present invention.

FIG. 3 is a diagram illustrating an example (a) of a motion compensation coding table according to the present invention.

FIG. 4 is a diagram illustrating an example (b) of a motion compensation coding table according to the present invention.

FIG. 5 is a diagram illustrating an example (c) of a motion compensation coding table according to the present invention.

FIG. 6 is a diagram illustrating an example (d) of a motion compensation coding table according to the present invention.

FIG. 7 is a flowchart (1) showing an operation example in the embodiment shown in FIG. 2;

FIG. 8 is a flowchart (2) showing an operation example in the embodiment shown in FIG. 2;

FIG. 9 is a flowchart (1) showing another operation example in the embodiment shown in FIG. 2;

FIG. 10 is a flowchart (2) showing another operation example in the embodiment shown in FIG. 2;

FIG. 11 is a flowchart showing still another operation example in the embodiment shown in FIG. 2;

FIG. 12 is a diagram illustrating an example (e) of a motion compensation coding table according to the present invention.

FIG. 13 is a diagram showing a conventional encoding circuit.

FIG. 14 is a diagram illustrating matching of a motion vector.

[Explanation of symbols]

1 motion detector 2 ₁ to 2 _n motion compensation coding table 3 memory 4 calculation and determination unit 5 _1, 5 ₂ switching unit

────────────────────────────────────────────────── ─── Continued on the front page (58) Field surveyed (Int.Cl. ⁶ , DB name) H04N ^7/ 24-7/68

Claims (4)

(57) [Claims] An encoded signal obtained by encoding a difference between a reproduced image immediately before an encoded frame and a prediction value is subjected to variable-length encoding, and a motion detecting unit (1) compares the encoded image signal with a reproduced image immediately before the encoded frame with an encoded signal. in high-efficiency video coding scheme for variable length coding using motion compensation coding table the detected motion vector by performing matching between the original image, a plurality of the motion compensation coding table (2 ₁ to 2 _n ), a memory (3) for holding the motion vector of the preceding one encoded frame, and a calculation / judgment unit (j) for judging the nature of the frequency of occurrence of the motion vector from the retained motion vector of the preceding encoded frame ( and 4), the switching unit for selecting the plurality of motion compensation coding table (2 ₁ to 2 _n) in accordance with the determination result (5 _1, 5 ₂₎ and is characterized in that the provided motion compensation variable Encoding method. 2. A calculation / determination unit (4) for determining whether or not the number of motion vectors that become 0 in one frame exceeds a certain threshold value, and a maximum value of the motion vectors in one frame. And generating information for selecting the plurality of motion compensation coding tables (2 ₁ to 2 _n ) in accordance with a result of determination as to whether or not the minimum value has exceeded each threshold value. 2. The motion-compensated variable-length coding method according to 1. 3. A determination result as to whether the number of motion vectors having the maximum value or the minimum value in one frame exceeds a respective threshold value,
Generating information for selecting the plurality of motion compensation coding tables (2 ₁ to 2 _n ) in accordance with a result of determining whether or not the number of motion vectors that become 0 in a frame exceeds a certain threshold value; The motion compensation variable length coding method according to claim 1, wherein: 4. The calculation / determination section (4) obtains an average value of motion vectors in one frame, and according to the average value, calculates the plurality of motion compensation coding tables (2 ₁ to 2 ₁ ).
The motion compensation variable length coding method according to claim 1, wherein information for selecting _2n ) is generated.